Technical interview questions for a CFD engineer : CFD

Hi all, I was hoping that we could share some technical questions we’ve been asked during interviews. I have one coming up and wanted to see different types of questions people have asked. I got my undergrad in aerospace engineering and my masters with a focus on CFD/aerodynamics. The role calls for someone familiar with flight sciences but it has been years since I’ve went in depth with subjects like flight dynamics.

Interview Questions of Top Designations

basic cfd interview questions

basic cfd interview questions

However, from my experience implementation of new (for the company, not newly developed) software techniques are closely tied with the availability of computational resources in the form of faster computers or solving on more CPUs. These are certainly ways to expand Computational Fluid Dynamics. Nevertheless, those aspects – better software and more application of Computational Fluid Dynamics – both really only address efficiency. At some point additional manpower must be addressed to realize any significant gains in output. That means hiring. And that “buck” stops at the manager – me[1]. However,while I was the CFD Group Leader I could completely advocate for additional CFD manpower whole heartily (and I did), but I must now take the interests of the entire department into account. That means balancing testing and computational resources based on a number of factors, such as immediate need or future need. First and foremost, from the company perspective, manpower is simply another resource that all departments are vying for (along with test facilities, equipment, test vehicles, computers, etc.).

Nevertheless, once authorization for additional manpower is agreed to, that means hiring, which in turn, means interviewing. I have been in a number of interviews – on both sides – and feel like I have a good idea of what I am looking for and what would work for our company. First and foremost, I am looking for someone that is interested in and enthusiastic about working in the automotive industry. Someone that has done their homework on our company. Someone that knows the fundamentals of their specialty. For CFD specifically, that means not only the rough/general equational development behind CFD, but also knowing and being able to explain aerodynamics and/or heat transfer. In my opinion, if you do not have some understanding of the physics or mathematics behind Computational Fluid Dynamics, then you will be less successful. I do not expect undergraduates to know the advantages/disadvantages of the various predictor-corrector numerical schemes or the advantages of finite volume techniques over finite difference techniques. In addition to any relevant work-related experience, I would (and do), however, ask the following question, as we interview a number of aerospace students:

These are not intended as “trick” questions as the interviewees are certainly welcome to ask clarifying questions, but I want to see them go through the process of solving the problem. I find this is invaluable to helping determining how someone will perform. Finally, specifically for CFD analysts, I ask if they are OK with sitting in at a computer workstation for 90-95% of the time. This is actually a deal breaker for a CFD position as I have found no matter how smart and/or capable a candidate is, if they cannot “sit still” or really want to “work with their hands”,they will not be successful at CFD.

Now that I have a few months to further settle in to the aero-thermal department manager role, I have had the chance to think more about expanding the role of Computational Fluid Dynamics within the department and potentially within the rest of the company as well. Our department certainly has an overall CFD strategy, but how that CFD strategy relates to the testing strategy within the entire development schedule could easily be re-thought. There is also the possibility of evaluating the always-developing software tools that we in the CAE community are well aware of.

You see, I want to know that they’ve considered the difference between the relatively balanced viscous and pressure forces on an airplane that they studied in college and the pressure-dominated forces on a road vehicle. I also expect them to understand problem solving as vehicle development is primarily about problem solving. To that end, I typically ask one seemingly innocuous question to see how they think. A couple of my favorites are:

10. Solutions of a system of PDEs can be obtained by equating the numerator of Cramer’s solution while using Cramer’s rule. This method is used by __________ a) Integral transform b) Change of variables c) Separation of variables d) Method of characteristics View AnswerAnswer: d Explanation: The Cramer’s numerator is equated to zero and compatibility equations are obtained. These are then solved to get the solution of PDEs.

4. What are the Cramer’s solutions equated to while using Cramer’s method of classifying a PDE? a) The dependent variables b) The derivatives of dependent variables c) The second derivatives of dependent variables d) The highest derivatives of dependent variables View AnswerAnswer: b Explanation: For characteristic lines, the derivatives of dependent variables are zero. Cramer’s rule is used to find these derivatives and then it is made indeterminate to find the type of PDE.

9. When the Eigenvalues are a mixture of real and imaginary values, the PDE is ___________ a) elliptic-hyperbolic b) parabolic c) elliptic d) hyperbolic View AnswerAnswer: a Explanation: When we get a mixed type of Eigenvalues, the type of PDE is also mixed. Many practical equations have mixed behaviour also.

7. Consider the following system of PDEs. (a_1frac{partial u}{partial x}+b_1frac{partial u}{partial y}+c_1frac{partial v}{partial x}+d_1frac{partial v}{partial y}=0) (a_2frac{partial u}{partial x}+b_2frac{partial u}{partial y}+c_2frac{partial v}{partial x}+d_2frac{partial v}{partial y}=0) The Eigenvalues of which of these matrices can be used to classify this system of PDEs? a) (begin{bmatrix} a_1 & b_1 a_2 & b_2 end{bmatrix}^{-1})(begin{bmatrix} c_1 & d_1 c_2 & d_2 end{bmatrix}) b) (begin{bmatrix} a_1 & c_1 a_2 & c_2 end{bmatrix}^{-1})(begin{bmatrix} b_1 & d_1 b_2 & d_2 end{bmatrix}) c) (begin{bmatrix} a_1 & c_1 a_2 & c_2 end{bmatrix})(begin{bmatrix} b_1 & d_1 b_2 & d_2 end{bmatrix}) d) (begin{bmatrix} a_1 & b_1 a_2 & b_2 end{bmatrix})(begin{bmatrix} c_1 & d_1 c_2 & d_2 end{bmatrix}) View AnswerAnswer: b Explanation: From the given system of PDEs, (a_1frac{partial u}{partial x}+b_1frac{partial u}{partial y}+c_1frac{partial v}{partial x}+d_1frac{partial v}{partial y}=0) (a_2frac{partial u}{partial x}+b_2frac{partial u}{partial y}+c_2frac{partial v}{partial x}+d_2frac{partial v}{partial y}=0) Let, (W=begin{bmatrix} u v end{bmatrix}) then (begin{bmatrix} a_1 & c_1 a_2 & c_2 end{bmatrix} frac{partial W}{partial x}+ begin{bmatrix} b_1 & d_1 b_2 & d_2 end{bmatrix} frac{partial W}{partial y}=0 ) (frac{partial W}{partial x}+ begin{bmatrix} a_1 & c_1 a_2 & c_2 end{bmatrix}^{-1} begin{bmatrix} b_1 & d_1 b_2 & d_2 end{bmatrix} frac{partial W}{partial y}=0) The Eigenvalues of (begin{bmatrix} a_1 & c_1 a_2 & c_2 end{bmatrix}^{-1})(begin{bmatrix} b_1 & d_1 b_2 & d_2 end{bmatrix}) determines the class of PDE.

3. How the type of PDE is identified using Cramer’s rule? a) By equating the Cramer’s denominator to 1 b) By equating the Cramer’s numerator to 1 c) By equating the Cramer’s denominator to 0 d) By equating the Cramer’s numerator to 0 View AnswerAnswer: c Explanation: The denominator of Cramer’s solution is equated to zero to find the type of PDE. The denominator is equated to zero to make the solution indeterminate.


Which method is used in CFD?

Essential Subjects for CFD Modeling
  • Mathematics: Partial differential equations, integration. …
  • Flow Physics: Fluid mechanics: Properties of fluid, Laminar, and Turbulent Flow. …
  • Computer science: Programming tools like C or C++ need to be learned by CFD users initially. …
  • Experimental data for validation:

What are the three stages of a CFD software?

The finite volume method (FVM) is a common approach used in CFD codes, as it has an advantage in memory usage and solution speed, especially for large problems, high Reynolds number turbulent flows, and source term dominated flows (like combustion).

Related Posts

Leave a Reply

Your email address will not be published.